OpenCV  4.2.0
Open Source Computer Vision
Public Member Functions | List of all members
cv::ximgproc::RICInterpolator Class Referenceabstract

Sparse match interpolation algorithm based on modified piecewise locally-weighted affine estimator called Robust Interpolation method of Correspondences or RIC from [106] and Variational and Fast Global Smoother as post-processing filter. The RICInterpolator is a extension of the EdgeAwareInterpolator. Main concept of this extension is an piece-wise affine model based on over-segmentation via SLIC superpixel estimation. The method contains an efficient propagation mechanism to estimate among the pieces-wise models. More...

#include <opencv2/ximgproc/sparse_match_interpolator.hpp>

Inheritance diagram for cv::ximgproc::RICInterpolator:
cv::ximgproc::SparseMatchInterpolator cv::Algorithm

Public Member Functions

virtual float getAlpha () const =0
 Alpha is a parameter defining a global weight for transforming geodesic distance into weight. More...
 
virtual float getFGSLambda () const =0
 Sets the respective fastGlobalSmootherFilter() parameter. More...
 
virtual float getFGSSigma () const =0
 Sets the respective fastGlobalSmootherFilter() parameter. More...
 
virtual int getK () const =0
 K is a number of nearest-neighbor matches considered, when fitting a locally affine model for a superpixel segment. However, lower values would make the interpolation noticeably faster. The original implementation of [106] uses 32. More...
 
virtual float getMaxFlow () const =0
 MaxFlow is a threshold to validate the predictions using a certain piece-wise affine model. If the prediction exceeds the treshold the translational model will be applied instead. More...
 
virtual int getModelIter () const =0
 Parameter defining the number of iterations for piece-wise affine model estimation. More...
 
virtual bool getRefineModels () const =0
 Parameter to choose wether additional refinement of the piece-wise affine models is employed. More...
 
virtual int getSuperpixelMode () const =0
 Parameter to choose superpixel algorithm variant to use: More...
 
virtual int getSuperpixelNNCnt () const =0
 Parameter defines the number of nearest-neighbor matches for each superpixel considered, when fitting a locally affine model. More...
 
virtual float getSuperpixelRuler () const =0
 Parameter to tune enforcement of superpixel smoothness factor used for oversegmentation. More...
 
virtual int getSuperpixelSize () const =0
 Get the internal cost, i.e. edge map, used for estimating the edge-aware term. More...
 
virtual bool getUseGlobalSmootherFilter () const =0
 Sets whether the fastGlobalSmootherFilter() post-processing is employed. More...
 
virtual bool getUseVariationalRefinement () const =0
 Parameter to choose wether the VariationalRefinement post-processing is employed. More...
 
virtual void setAlpha (float alpha=0.7f)=0
 Alpha is a parameter defining a global weight for transforming geodesic distance into weight. More...
 
virtual void setCostMap (const Mat &costMap)=0
 Interface to provide a more elaborated cost map, i.e. edge map, for the edge-aware term. This implementation is based on a rather simple gradient-based edge map estimation. To used more complex edge map estimator (e.g. StructuredEdgeDetection that has been used in the original publication) that may lead to improved accuracies, the internal edge map estimation can be bypassed here. More...
 
virtual void setFGSLambda (float lambda=500.f)=0
 Sets the respective fastGlobalSmootherFilter() parameter. More...
 
virtual void setFGSSigma (float sigma=1.5f)=0
 Sets the respective fastGlobalSmootherFilter() parameter. More...
 
virtual void setK (int k=32)=0
 K is a number of nearest-neighbor matches considered, when fitting a locally affine model for a superpixel segment. However, lower values would make the interpolation noticeably faster. The original implementation of [106] uses 32. More...
 
virtual void setMaxFlow (float maxFlow=250.f)=0
 MaxFlow is a threshold to validate the predictions using a certain piece-wise affine model. If the prediction exceeds the treshold the translational model will be applied instead. More...
 
virtual void setModelIter (int modelIter=4)=0
 Parameter defining the number of iterations for piece-wise affine model estimation. More...
 
virtual void setRefineModels (bool refineModles=true)=0
 Parameter to choose wether additional refinement of the piece-wise affine models is employed. More...
 
virtual void setSuperpixelMode (int mode=100)=0
 Parameter to choose superpixel algorithm variant to use: More...
 
virtual void setSuperpixelNNCnt (int spNN=150)=0
 Parameter defines the number of nearest-neighbor matches for each superpixel considered, when fitting a locally affine model. More...
 
virtual void setSuperpixelRuler (float ruler=15.f)=0
 Parameter to tune enforcement of superpixel smoothness factor used for oversegmentation. More...
 
virtual void setSuperpixelSize (int spSize=15)=0
 Get the internal cost, i.e. edge map, used for estimating the edge-aware term. More...
 
virtual void setUseGlobalSmootherFilter (bool use_FGS=true)=0
 Sets whether the fastGlobalSmootherFilter() post-processing is employed. More...
 
virtual void setUseVariationalRefinement (bool use_variational_refinement=false)=0
 Parameter to choose wether the VariationalRefinement post-processing is employed. More...
 
- Public Member Functions inherited from cv::ximgproc::SparseMatchInterpolator
virtual void interpolate (InputArray from_image, InputArray from_points, InputArray to_image, InputArray to_points, OutputArray dense_flow)=0
 Interpolate input sparse matches. More...
 
- Public Member Functions inherited from cv::Algorithm
 Algorithm ()
 
virtual ~Algorithm ()
 
virtual void clear ()
 Clears the algorithm state. More...
 
virtual bool empty () const
 Returns true if the Algorithm is empty (e.g. in the very beginning or after unsuccessful read. More...
 
virtual String getDefaultName () const
 
virtual void read (const FileNode &fn)
 Reads algorithm parameters from a file storage. More...
 
virtual void save (const String &filename) const
 
virtual void write (FileStorage &fs) const
 Stores algorithm parameters in a file storage. More...
 
void write (const Ptr< FileStorage > &fs, const String &name=String()) const
 simplified API for language bindings This is an overloaded member function, provided for convenience. It differs from the above function only in what argument(s) it accepts. More...
 

Additional Inherited Members

- Static Public Member Functions inherited from cv::Algorithm
template<typename _Tp >
static Ptr< _Tp > load (const String &filename, const String &objname=String())
 Loads algorithm from the file. More...
 
template<typename _Tp >
static Ptr< _Tp > loadFromString (const String &strModel, const String &objname=String())
 Loads algorithm from a String. More...
 
template<typename _Tp >
static Ptr< _Tp > read (const FileNode &fn)
 Reads algorithm from the file node. More...
 
- Protected Member Functions inherited from cv::Algorithm
void writeFormat (FileStorage &fs) const
 

Detailed Description

Sparse match interpolation algorithm based on modified piecewise locally-weighted affine estimator called Robust Interpolation method of Correspondences or RIC from [106] and Variational and Fast Global Smoother as post-processing filter. The RICInterpolator is a extension of the EdgeAwareInterpolator. Main concept of this extension is an piece-wise affine model based on over-segmentation via SLIC superpixel estimation. The method contains an efficient propagation mechanism to estimate among the pieces-wise models.

Member Function Documentation

◆ getAlpha()

virtual float cv::ximgproc::RICInterpolator::getAlpha ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getAlpha()

Alpha is a parameter defining a global weight for transforming geodesic distance into weight.

See also
setAlpha

◆ getFGSLambda()

virtual float cv::ximgproc::RICInterpolator::getFGSLambda ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getFGSLambda()

Sets the respective fastGlobalSmootherFilter() parameter.

See also
setFGSLambda

◆ getFGSSigma()

virtual float cv::ximgproc::RICInterpolator::getFGSSigma ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getFGSSigma()

Sets the respective fastGlobalSmootherFilter() parameter.

See also
setFGSSigma

◆ getK()

virtual int cv::ximgproc::RICInterpolator::getK ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getK()

K is a number of nearest-neighbor matches considered, when fitting a locally affine model for a superpixel segment. However, lower values would make the interpolation noticeably faster. The original implementation of [106] uses 32.

See also
setK

◆ getMaxFlow()

virtual float cv::ximgproc::RICInterpolator::getMaxFlow ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getMaxFlow()

MaxFlow is a threshold to validate the predictions using a certain piece-wise affine model. If the prediction exceeds the treshold the translational model will be applied instead.

See also
setMaxFlow

◆ getModelIter()

virtual int cv::ximgproc::RICInterpolator::getModelIter ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getModelIter()

Parameter defining the number of iterations for piece-wise affine model estimation.

See also
setModelIter

◆ getRefineModels()

virtual bool cv::ximgproc::RICInterpolator::getRefineModels ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getRefineModels()

Parameter to choose wether additional refinement of the piece-wise affine models is employed.

See also
setRefineModels

◆ getSuperpixelMode()

virtual int cv::ximgproc::RICInterpolator::getSuperpixelMode ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getSuperpixelMode()

Parameter to choose superpixel algorithm variant to use:

See also
setSuperpixelMode

◆ getSuperpixelNNCnt()

virtual int cv::ximgproc::RICInterpolator::getSuperpixelNNCnt ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getSuperpixelNNCnt()

Parameter defines the number of nearest-neighbor matches for each superpixel considered, when fitting a locally affine model.

See also
setSuperpixelNNCnt

◆ getSuperpixelRuler()

virtual float cv::ximgproc::RICInterpolator::getSuperpixelRuler ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getSuperpixelRuler()

Parameter to tune enforcement of superpixel smoothness factor used for oversegmentation.

See also
setSuperpixelRuler

◆ getSuperpixelSize()

virtual int cv::ximgproc::RICInterpolator::getSuperpixelSize ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getSuperpixelSize()

Get the internal cost, i.e. edge map, used for estimating the edge-aware term.

See also
setSuperpixelSize

◆ getUseGlobalSmootherFilter()

virtual bool cv::ximgproc::RICInterpolator::getUseGlobalSmootherFilter ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getUseGlobalSmootherFilter()

Sets whether the fastGlobalSmootherFilter() post-processing is employed.

See also
setUseGlobalSmootherFilter

◆ getUseVariationalRefinement()

virtual bool cv::ximgproc::RICInterpolator::getUseVariationalRefinement ( ) const
pure virtual
Python:
retval=cv.ximgproc_RICInterpolator.getUseVariationalRefinement()

Parameter to choose wether the VariationalRefinement post-processing is employed.

See also
setUseVariationalRefinement

◆ setAlpha()

virtual void cv::ximgproc::RICInterpolator::setAlpha ( float  alpha = 0.7f)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setAlpha([, alpha])

Alpha is a parameter defining a global weight for transforming geodesic distance into weight.

◆ setCostMap()

virtual void cv::ximgproc::RICInterpolator::setCostMap ( const Mat costMap)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setCostMap(costMap)

Interface to provide a more elaborated cost map, i.e. edge map, for the edge-aware term. This implementation is based on a rather simple gradient-based edge map estimation. To used more complex edge map estimator (e.g. StructuredEdgeDetection that has been used in the original publication) that may lead to improved accuracies, the internal edge map estimation can be bypassed here.

Parameters
costMapa type CV_32FC1 Mat is required.
See also
cv::ximgproc::createSuperpixelSLIC

◆ setFGSLambda()

virtual void cv::ximgproc::RICInterpolator::setFGSLambda ( float  lambda = 500.f)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setFGSLambda([, lambda])

Sets the respective fastGlobalSmootherFilter() parameter.

◆ setFGSSigma()

virtual void cv::ximgproc::RICInterpolator::setFGSSigma ( float  sigma = 1.5f)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setFGSSigma([, sigma])

Sets the respective fastGlobalSmootherFilter() parameter.

◆ setK()

virtual void cv::ximgproc::RICInterpolator::setK ( int  k = 32)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setK([, k])

K is a number of nearest-neighbor matches considered, when fitting a locally affine model for a superpixel segment. However, lower values would make the interpolation noticeably faster. The original implementation of [106] uses 32.

◆ setMaxFlow()

virtual void cv::ximgproc::RICInterpolator::setMaxFlow ( float  maxFlow = 250.f)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setMaxFlow([, maxFlow])

MaxFlow is a threshold to validate the predictions using a certain piece-wise affine model. If the prediction exceeds the treshold the translational model will be applied instead.

◆ setModelIter()

virtual void cv::ximgproc::RICInterpolator::setModelIter ( int  modelIter = 4)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setModelIter([, modelIter])

Parameter defining the number of iterations for piece-wise affine model estimation.

◆ setRefineModels()

virtual void cv::ximgproc::RICInterpolator::setRefineModels ( bool  refineModles = true)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setRefineModels([, refineModles])

Parameter to choose wether additional refinement of the piece-wise affine models is employed.

◆ setSuperpixelMode()

virtual void cv::ximgproc::RICInterpolator::setSuperpixelMode ( int  mode = 100)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setSuperpixelMode([, mode])

Parameter to choose superpixel algorithm variant to use:

◆ setSuperpixelNNCnt()

virtual void cv::ximgproc::RICInterpolator::setSuperpixelNNCnt ( int  spNN = 150)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setSuperpixelNNCnt([, spNN])

Parameter defines the number of nearest-neighbor matches for each superpixel considered, when fitting a locally affine model.

◆ setSuperpixelRuler()

virtual void cv::ximgproc::RICInterpolator::setSuperpixelRuler ( float  ruler = 15.f)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setSuperpixelRuler([, ruler])

Parameter to tune enforcement of superpixel smoothness factor used for oversegmentation.

See also
cv::ximgproc::createSuperpixelSLIC

◆ setSuperpixelSize()

virtual void cv::ximgproc::RICInterpolator::setSuperpixelSize ( int  spSize = 15)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setSuperpixelSize([, spSize])

Get the internal cost, i.e. edge map, used for estimating the edge-aware term.

See also
setCostMap

◆ setUseGlobalSmootherFilter()

virtual void cv::ximgproc::RICInterpolator::setUseGlobalSmootherFilter ( bool  use_FGS = true)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setUseGlobalSmootherFilter([, use_FGS])

Sets whether the fastGlobalSmootherFilter() post-processing is employed.

◆ setUseVariationalRefinement()

virtual void cv::ximgproc::RICInterpolator::setUseVariationalRefinement ( bool  use_variational_refinement = false)
pure virtual
Python:
None=cv.ximgproc_RICInterpolator.setUseVariationalRefinement([, use_variational_refinement])

Parameter to choose wether the VariationalRefinement post-processing is employed.


The documentation for this class was generated from the following file: